XClose

UCL Psychology and Language Sciences

Home
Menu

Agreement Mismatches in Experimental Syntax: from Slavic to Bantu

The project examines combinatorial possibilities of gender agreement in coordinations, coupled with challenges of modeling animacy effects and the structure of coordination, providing support from two typologically unrelated families, South Slavic and Southern Bantu, for linear agreement as represented in grammar and language processing. The originality of the project lies in its idea of crossfamily comparison across two apparently very unrelated families that nonetheless share complex systems of gender replete with mismatches and observed microvariation in agreement strategies, and typological comparison through previously established experimental design paradigms. Therefore, this project takes a radically new approach to understanding how languages vary by applying methodology established by previous groundbreaking work rolled out across language groups (rather than within): a theoretically-informed comparison of experimental syntax fieldwork, conducted with research partners from the former Yugoslavia and South Africa.


Funding Source: Leverhulme Trust (2018 - 2022)  

Principal Investigator: Prof Andrew Nevins

Postdoctoral Research Associate: Dr Jana Willer-Gold

Co-Investigators: Prof Diogo Almeida, Prof Mark de Vos, Prof Franc Marušič, Prof Anita Peti-Stantić, Prof Ron Simango, Prof Jochen Zeller, Dr Kristina Riedel, Ms Hazel Mitchley 



CONTACT

Email:

a.nevins@ucl.ac.uk or j.willer-gold@ucl.ac.uk

Telephone:

+ 44 (0) 20 7679 4278 (x24278)

Address:

Chandler House 
2 Wakefield Street 
London 
WC1N 1PF


RESEARCH 

Agreement has taken on an increasingly important role in shaping grammatical theory, providing evidence about constraints on structural, morphosyntactic, and other relations (e.g., interactions between agreement and case; anti-agreement; and the syntactic locality and scope of agreement operations, e.g. Chung 1998), about the functional properties or projections of phrasal heads (e.g., Baker 2008), and about the interface between syntax and morphology (e.g., whether agreement is confined to syntactic or morphological components or is distributed across multiple components of the grammar; e.g., Bobaljik 2008). In this project, we address three issues that are of fundamental interest to the theory of agreement:

  1. What role do syntactic hierarchy, linear contiguity, and featural markedness play in the calculation of agreement (either in general, or in the special circumstances that conjoined phrases create)? 
  2. How does the agreement function with regard to the unity versus independence of inflectional features (e.g., do the mechanisms that probe arguments for number and gender features operate jointly or independently, and are these alternatives mutually exclusive for a language)? 
  3. To what degree is syntactic variability attributable to how grammars strike a balance between competing grammatical mechanisms?

Our primary focus is on coordinated noun phrases. The structure of coordinated phrases has a long history in the language sciences, where in the philosophy of language and semantic theory, conjunction is treated as a symmetric operator (Montague 1973, Partee & Rooth 1983), whereas research in syntactic theory has developed extensive evidence that the internal structure of coordinations is asymmetric and hierarchical (Goodall 1983, Munn 1992, Johannessen 1993, Kayne 1994, De Vos 2009). Given this latter structure and the scarcity of purely linear-order relations within syntax, it is expected that the two conjoined noun phrases in a coordination would not be equally possible agreement controllers. Yet a number of studies have recently argued that linear order is a relevant relation for syntactic operations specifically when it comes to coordinated phrases, as they are headed by neither one of the conjuncts (Marušic, Nevins & Saksida 2007, Bhatt & Walkow 2013). Coordinations therefore offer an opportunity to examine whether agreement morphology may operate with its own principles, partially distinct from those of other syntactic relations, where the norm is reliance on hierarchical relations.

Therefore, studying South Slavic and Southern Bantu languages in their local contexts is of particularly interest given:

  • The importance of theoretically-informed experimental linguistics research on local languages in a highly multilingual society. 
  • The extremely rich morphosyntactic system of noun classes found in these languages, unparalleled within European languages that form the primary basis of current theoretical models of conjunct agreement. 
  • The clearly articulated grammatical distinction between [+/- human] noun phrases that leads to different systems of defaults and indeed to bans on mixed humanness, and (d) the theoretical challenges posed by disambiguating between coordination and comitative constructions that are often syncretic and therefore designing careful syntactic diagnostics.

METHODOLOGY

Experimental syntax is a burgeoning field that provides methodologies to study the processing of linguistically challenging structures that present variation within and across languages, and moreover guarantees a more robust empirical foundation for resulting linguistic theories. Such theoretically-informed psycholinguistic studies of gender and number agreement in complex configurations have seen increased interest in recent years (Haskell & MacDonald 2005, Franck et al 2007, Keung & Staub 2017), but such techniques have not yet been applied to the rich morphological combinatorics of the Bantu languages. Given the complexity and challenges of experimental design on South Slavic and Bantu family of languages and the extreme combinatorics of gender and number within coordinations, a thorough suite of experiments will require a factorial design extending over many participant groups. There is clear importance for research on Southern Bantu languages within the context of South Africa and their importance they bear for linguistic theory itself, with the potential to transform existing theories of the structure of conjunction, the representation of humanness and animacy in the grammar, and the nature of featural representation in complex number/gender systems. Existing reports on conjunct agreement in Southern Bantu have not yet employed such methods, which our research team has already shown a proven track record in its application to South Slavic. In parallel, given that the role of mixed humanness, strikingly clear in Southern Bantu, has never been experimentally studied in South Slavic, we aim to carry out experiments of this sort in Slovenian and Bosnian/Croatian/Serbian -- another case in which, alongside the comparison of coordination and comitatives in Slavic (Vassilieva & Larson 2001). All of these theoretical developments in turn can feed back into more refined research questions applied to South Slavic, and to the fostering of dialogue between researchers within these two regions.

Experimental syntax methodology applied in our research includes the following elements (cf. Schütze 1996, Cowart 1997, Featherston 2007, Goodall 2011): 

  • Nonlinguist study participants: We will be recruiting university students who are not studying language/linguistics. 
  • Clearly defined tasks: We will employ an elicited production task where participants will be asked to create a sentence by inserting a novel subject noun phrase into a model sentence and thus figuring by themselves what agreement should be used on the verb (early measures), and a timed acceptability judgment task in which participants evaluate the naturalness of a given sentence (late measures). 
  • Factorial design for the construction of stimuli: Research will involve elicited production across a suite of combinations, akin to those examined in Marušic et al 2015, Willer Gold et al 2016, and Willer Gold et al 2018 for South Slavic. In each of the planned experiments, the tested hypotheses will be broken down to individual factors which will be tested against controls, with counterbalanced stimuli and equal numbers of items for each variable of interest. Only with such a design one can make a strong claim about a particular phenomenon.
  • Quantitative results: Statistical analysis using mixed models effects of both production responses and overall production latency can reveal the effects of the number and gender of each conjunct, word order effects, participant and item effects, and the interactions of all of these main effects. Principal Component Analysis can be used across language varieties to look for clustering effects and the construction of an experimentally elicited morphosyntactic atlas of language distances (as performed in Willer Gold et al 2018).

OBJECTIVES

  • To produce a robust set of empirical results, using experimental syntax methodology, on conjunct agreement in Southern Bantu through large-scale studies with dozens of participants from each language sampled.

  • To train and engage a new generation of partners in South Africa and to forge research links between the UK, the former Yugoslavia, and South Africa that can apply such partnerships and methodologies to a host of new research questions.

  • To confirm and modify the theory of agreement controllers and the choice among strategies of highest conjunct agreement, closest conjunct agreement, default agreement, and resolved agreement and their distribution within and across Southern Bantu.

  • To refine the theory of the representation of gender, humanness, and defaults in noun classes in Bantu in terms of an articulated set of features and their distribution on agreement controllers and targets.


RELEVANT PUBLICATIONS

  • Corbett, Greville G. and Alfred D. Mtenje. 1987. Gender agreement in Chichewa. Studies in African Linguistics 18.1: 1–38.
  • de Vos, Mark and Hazel Mitchley. 2012. Subject marking and preverbal coordination in Sesotho: A perspective from Optimality Theory. Southern African Linguistics and Applied Language Studies 30.2: 155-170.
  • Nevins, Andrew. 2018. Copying and resolution in South Slavic and South Bantu Conjunct Agreement. In Petrosino, Roberto, Pietro Cerrone and Harry van der Hulst (eds.)  From Sounds to Structures Beyond the Veil of Maya. Series: Studies in Generative Grammar 135, De Gruyter. 391-408. 
  • Willer Gold Jana, Boban Arsenijević, Mia Batinić, Nermina Čordalija, Michael Becker, Marijana Kresić, Nedžad Leko, Franc Lanko Marušič, Tanja Milićev, Nataša Milićević, Ivana Mitić, Anita Peti-Stantić, Branimir Stanković, Tina Šuligoj, Jelena Tušek and Andrew Nevins. 2018. When Linearity Prevails over Hierarchy in Syntax. Proceedings of the National Academy of Sciences 115.3: 495-500.
  • Willer Gold Jana, Boban Arsenijević, Mia Batinić, Nermina Cordalija, Marijana Kresić, Nedžad Leko, Franc Lanko Marušič, Tanja Milićev, Nataša Milićević, Ivana Mitić, Anita Peti-Stantić, Branimir Stanković, Tina Šuligoj, Jelena Tušek and Andrew Nevins. 2016. Conjunct Agreement and Gender in South Slavic. Journal of Slavic Linguistics 24.1: 187-224.
  • Boban Arsenijević, Jana Willer-Gold, Nadira Aljović, Nermina Cordalija, Marijana Kresić Vukosav, Nedzad Leko, Frane Malenica, Franc Lanko Marusic, Tanja Milicev, Natasa Milicević, Petra Mismas, Ivana Mitić, Anita Peti-Stantić, Branimir Stanković, Jelena Tusek, and Andrew Nevins. 2019. Elided Clausal Conjunction Is Not the Only Source of Closest‐Conjunct Agreement: A Picture‐Matching StudySyntax